From mboxrd@z Thu Jan 1 00:00:00 1970 From: Bill Fink Subject: Re: [PATCH] net: add ability to clear stats via ethtool - e1000/pcnet32 Date: Sat, 31 May 2008 00:47:03 -0400 Message-ID: <20080531004703.a081c18f.billfink@mindspring.com> References: <482DA5B6.1020606@sngx.net> <482DB46A.8020103@cosmosbay.com> <482EF192.4070707@sngx.net> <482F5113.5090703@cosmosbay.com> <482F610D.2080108@sngx.net> <20080518003104.GK28241@solarflare.com> <482FBA09.80201@sngx.net> <483E0AAE.2020107@sngx.net> <20080528221118.63da4092.akpm@linux-foundation.org> <483EA2D1.8050603@sngx.net> <20080529154525.3916c7b5@core> <20080530151250.b44a119a.billfink@mindspring.com> <48407C51.7090405@hp.com> <20080530210931.5970c975.billfink@mindspring.com> <20080530194113.25b77988@extreme> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Cc: netdev@vger.kernel.org, linux-kernel@vger.kernel.org To: Stephen Hemminger Return-path: Received: from elasmtp-kukur.atl.sa.earthlink.net ([209.86.89.65]:40964 "EHLO elasmtp-kukur.atl.sa.earthlink.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750889AbYEaErJ (ORCPT ); Sat, 31 May 2008 00:47:09 -0400 In-Reply-To: <20080530194113.25b77988@extreme> Sender: netdev-owner@vger.kernel.org List-ID: On Fri, 30 May 2008, Stephen Hemminger wrote: > On Fri, 30 May 2008 21:09:31 -0400 > Bill Fink wrote: > > > On Fri, 30 May 2008, Rick Jones wrote: > > > > > > Saving stats output and running beforeafter on a number of systems is > > > > a royal pain when troubleshooting. > > > > > > Well, with that ringing endorsement :) I would like to state that > > > Yoshihiro Ishijima has fixed some bugs in beforeafter and it will now > > > deal with /proc/interrupts output better than before. > > > > > > ftp://ftp.cup.hp.com/dist/networking/tools/ > > > > Nothing against beforefter. It's a very useful tool, especially > > given Linux's lack of a method to clear interface counters. But > > it still is very cumbersome to use in the scenario I mentioned, > > having to save the stats before and after a test on each of several > > systems, and then having to run beforeafter on each of those systems. > > Then you need to do this all again for every test run. It's just > > not very efficient. > > > > I don't know why the major objection to adding a small amount of > > code to the kernel that would make life a lot easier for hundreds > > (or more) of Linux network adminstrators. And the snapshot/diff > > idea would not be complex to maintain and would also address the > > issue of some hardware not having a clear hw stats capability. > > > > -Bill > > ifstat from iproute2 tools also keeps state to compute number > of packets since last update as well. Come on, it isn't that > hard. Thanks for the pointer. I wasn't aware of ifstat (it's not installed by default on any of my systems). I will check out ifstat and see what it can do. -Thanks -Bill